During the current fiscal year we have determined optimal stimulus shapes for neuronal excitation using the squid giant axon as a model system. These experiments employed excitatory and inhibitory post-synaptic potentials delivered to the axon at computer-generated random times. The overall intensity of the stimulus was titrated to yield relatively infrequent action potentials, or spikes. The shapes of current input that elicited action potentials were determined from spike-triggered averaging. These results are optimal in the sense that minimal amounts of current were required to produce spikes. The work is relevant to deep brain stimulation (DBS) protocols used to treat movement disorders in patients with Parkinson's disease. The stimulation protocol typically used in DBS consists of rectangular current pulses. A noise based approach may be more efficacious not only in prolonging battery life of the battery pack implanted in these patients but also in a qualitative improvement in treatment of movement disorders.